{ "environment": { "iso_timestamp": "2026-09-03T12:31:47Z", "liboqs_version": "0.15.0", "liboqs_python_version": "0.15.0", "git_commit": "ca4dd54690c18c4564203569f981f3f0e8c0bb64", "cpu_model": "Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz", "arch": "x86_64", "build_path": "OQS_DIST_BUILD dispatch \u00b7 kyber768/falcon AVX2 \u00b7 SHA3 AVX512VL", "steal_time_pct": 6.0005 }, "families": { "post-quantum": [ "ML-DSA-44", "ML-DSA-65", "ML-DSA-87", "Falcon-512", "Falcon-1024", "SLH_DSA_PURE_SHAKE_128S", "SLH_DSA_PURE_SHAKE_128F" ], "classical": [ "RSA-2048-PSS", "RSA-3072-PSS", "ECDSA-P256", "ECDSA-P384" ] }, "comparison_note": "Classical and post-quantum schemes in this file were measured in the same run on the same host, so a delta between any two of them is a same-run comparison. No pairing between a classical scheme and a post-quantum one is asserted: that is a security-level argument rather than a measurement, and each classical record publishes its documented strength so the argument can be made explicitly.", "schemes": { "ML-DSA-44": { "scheme": "ML-DSA-44", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 37.212, "median_us": 32.386, "p95_us": 50.082, "p99_us": 132.941, "stdev_us": 21.486, "min_us": 31.069, "max_us": 356.858, "ops_per_sec": 26872.82, "n_iterations": 1000, "ci95_low_us": 35.881, "ci95_high_us": 38.544, "std_error_us": 0.6795, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 114.46, "median_us": 91.025, "p95_us": 254.576, "p99_us": 387.712, "stdev_us": 74.424, "min_us": 47.253, "max_us": 742.954, "ops_per_sec": 8736.68, "n_iterations": 1000, "ci95_low_us": 109.847, "ci95_high_us": 119.073, "std_error_us": 2.3535, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 35.817, "median_us": 32.633, "p95_us": 47.024, "p99_us": 122.369, "stdev_us": 14.818, "min_us": 32.212, "max_us": 226.803, "ops_per_sec": 27919.68, "n_iterations": 1000, "ci95_low_us": 34.899, "ci95_high_us": 36.735, "std_error_us": 0.4686, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 2420, "public_key_bytes": 1312, "secret_key_bytes": 2560 }, "ML-DSA-65": { "scheme": "ML-DSA-65", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 52.536, "median_us": 51.414, "p95_us": 58.48, "p99_us": 63.478, "stdev_us": 8.445, "min_us": 50.973, "max_us": 247.439, "ops_per_sec": 19034.58, "n_iterations": 1000, "ci95_low_us": 52.013, "ci95_high_us": 53.059, "std_error_us": 0.2671, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 170.563, "median_us": 133.7, "p95_us": 385.461, "p99_us": 661.304, "stdev_us": 116.86, "min_us": 68.328, "max_us": 946.587, "ops_per_sec": 5862.95, "n_iterations": 1000, "ci95_low_us": 163.32, "ci95_high_us": 177.806, "std_error_us": 3.6954, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 65.448, "median_us": 51.483, "p95_us": 98.934, "p99_us": 236.068, "stdev_us": 41.34, "min_us": 50.125, "max_us": 761.001, "ops_per_sec": 15279.21, "n_iterations": 1000, "ci95_low_us": 62.886, "ci95_high_us": 68.011, "std_error_us": 1.3073, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 3309, "public_key_bytes": 1952, "secret_key_bytes": 4032 }, "ML-DSA-87": { "scheme": "ML-DSA-87", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 87.15, "median_us": 77.025, "p95_us": 109.965, "p99_us": 285.506, "stdev_us": 42.703, "min_us": 73.615, "max_us": 661.326, "ops_per_sec": 11474.41, "n_iterations": 1000, "ci95_low_us": 84.504, "ci95_high_us": 89.797, "std_error_us": 1.3504, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 203.723, "median_us": 162.656, "p95_us": 448.563, "p99_us": 660.886, "stdev_us": 123.851, "min_us": 97.79, "max_us": 1114.307, "ops_per_sec": 4908.63, "n_iterations": 1000, "ci95_low_us": 196.047, "ci95_high_us": 211.399, "std_error_us": 3.9165, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 82.252, "median_us": 73.179, "p95_us": 103.531, "p99_us": 249.764, "stdev_us": 39.329, "min_us": 72.318, "max_us": 712.07, "ops_per_sec": 12157.71, "n_iterations": 1000, "ci95_low_us": 79.815, "ci95_high_us": 84.69, "std_error_us": 1.2437, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 4627, "public_key_bytes": 2592, "secret_key_bytes": 4896 }, "Falcon-512": { "scheme": "Falcon-512", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 11741.59, "median_us": 9845.171, "p95_us": 22758.651, "p99_us": 38921.011, "stdev_us": 6231.169, "min_us": 6625.001, "max_us": 77764.995, "ops_per_sec": 85.17, "n_iterations": 1000, "ci95_low_us": 11355.385, "ci95_high_us": 12127.795, "std_error_us": 197.0469, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 422.774, "median_us": 359.827, "p95_us": 677.884, "p99_us": 1105.441, "stdev_us": 173.714, "min_us": 336.989, "max_us": 2966.138, "ops_per_sec": 2365.33, "n_iterations": 1000, "ci95_low_us": 412.007, "ci95_high_us": 433.541, "std_error_us": 5.4933, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 82.597, "median_us": 73.012, "p95_us": 122.132, "p99_us": 226.267, "stdev_us": 34.965, "min_us": 70.533, "max_us": 567.182, "ops_per_sec": 12106.92, "n_iterations": 1000, "ci95_low_us": 80.43, "ci95_high_us": 84.765, "std_error_us": 1.1057, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 752, "public_key_bytes": 897, "secret_key_bytes": 1281 }, "Falcon-1024": { "scheme": "Falcon-1024", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 29644.258, "median_us": 27049.448, "p95_us": 46086.95, "p99_us": 58475.309, "stdev_us": 8086.699, "min_us": 20459.202, "max_us": 72593.151, "ops_per_sec": 33.73, "n_iterations": 1000, "ci95_low_us": 29143.049, "ci95_high_us": 30145.468, "std_error_us": 255.7239, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 836.657, "median_us": 723.553, "p95_us": 1278.98, "p99_us": 1744.06, "stdev_us": 235.086, "min_us": 657.389, "max_us": 2728.871, "ops_per_sec": 1195.23, "n_iterations": 1000, "ci95_low_us": 822.086, "ci95_high_us": 851.227, "std_error_us": 7.4341, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 160.602, "median_us": 140.994, "p95_us": 255.672, "p99_us": 480.293, "stdev_us": 72.843, "min_us": 133.948, "max_us": 1060.38, "ops_per_sec": 6226.58, "n_iterations": 1000, "ci95_low_us": 156.087, "ci95_high_us": 165.116, "std_error_us": 2.3035, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 1462, "public_key_bytes": 1793, "secret_key_bytes": 2305 }, "SLH_DSA_PURE_SHAKE_128S": { "scheme": "SLH_DSA_PURE_SHAKE_128S", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 190101.673, "median_us": 179731.567, "p95_us": 230447.59, "p99_us": 375786.317, "stdev_us": 42151.343, "min_us": 167780.795, "max_us": 773525.274, "ops_per_sec": 5.26, "n_iterations": 1000, "ci95_low_us": 187489.154, "ci95_high_us": 192714.192, "std_error_us": 1332.9425, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 1566827.651, "median_us": 1488777.166, "p95_us": 1957614.843, "p99_us": 2978391.514, "stdev_us": 299320.882, "min_us": 1318598.192, "max_us": 3713499.442, "ops_per_sec": 0.64, "n_iterations": 1000, "ci95_low_us": 1548275.892, "ci95_high_us": 1585379.411, "std_error_us": 9465.3574, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 1546.895, "median_us": 1369.331, "p95_us": 2308.926, "p99_us": 3565.237, "stdev_us": 775.082, "min_us": 1134.885, "max_us": 18696.283, "ops_per_sec": 646.46, "n_iterations": 1000, "ci95_low_us": 1498.856, "ci95_high_us": 1594.935, "std_error_us": 24.5103, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 7856, "public_key_bytes": 32, "secret_key_bytes": 64 }, "SLH_DSA_PURE_SHAKE_128F": { "scheme": "SLH_DSA_PURE_SHAKE_128F", "kind": "post-quantum", "status": "ok", "keygen": { "mean_us": 2758.05, "median_us": 2658.578, "p95_us": 3169.977, "p99_us": 4080.682, "stdev_us": 279.561, "min_us": 2597.791, "max_us": 5897.519, "ops_per_sec": 362.57, "n_iterations": 1000, "ci95_low_us": 2740.723, "ci95_high_us": 2775.377, "std_error_us": 8.8405, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 70827.478, "median_us": 65808.019, "p95_us": 87312.137, "p99_us": 168610.185, "stdev_us": 20580.871, "min_us": 61487.947, "max_us": 341763.727, "ops_per_sec": 14.12, "n_iterations": 1000, "ci95_low_us": 69551.885, "ci95_high_us": 72103.07, "std_error_us": 650.8243, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 4208.717, "median_us": 3900.465, "p95_us": 5699.54, "p99_us": 7441.347, "stdev_us": 916.03, "min_us": 3437.822, "max_us": 13781.702, "ops_per_sec": 237.6, "n_iterations": 1000, "ci95_low_us": 4151.942, "ci95_high_us": 4265.492, "std_error_us": 28.9674, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 17088, "public_key_bytes": 32, "secret_key_bytes": 64 }, "RSA-2048-PSS": { "scheme": "RSA-2048-PSS", "family": "RSA-PSS", "kind": "classical", "status": "ok", "keygen": { "mean_us": 70866.502, "median_us": 55074.797, "p95_us": 198285.727, "p99_us": 220571.846, "stdev_us": 54031.959, "min_us": 17023.316, "max_us": 226143.376, "ops_per_sec": 14.11, "n_iterations": 20, "ci95_low_us": 47186.387, "ci95_high_us": 94546.617, "std_error_us": 12081.9133, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 781.462, "median_us": 686.61, "p95_us": 1186.236, "p99_us": 1604.418, "stdev_us": 223.485, "min_us": 654.444, "max_us": 3421.109, "ops_per_sec": 1279.65, "n_iterations": 1000, "ci95_low_us": 767.61, "ci95_high_us": 795.313, "std_error_us": 7.0672, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 75.438, "median_us": 59.629, "p95_us": 175.159, "p99_us": 362.228, "stdev_us": 53.329, "min_us": 45.361, "max_us": 439.628, "ops_per_sec": 13255.9, "n_iterations": 1000, "ci95_low_us": 72.133, "ci95_high_us": 78.743, "std_error_us": 1.6864, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 256, "public_key_bytes": 294, "parameters": { "hash": "SHA-256", "padding": "PSS", "salt_length": "digest length (32 B)" }, "classical_security_bits": 112, "security_source": "NIST SP 800-57 Part 1 Rev. 5, Table 2", "pairing_note": "No post-quantum counterpart is asserted. Pairing a classical scheme with a post-quantum one is a security-level claim rather than a measurement; the documented strength is published so the pairing can be argued rather than assumed.", "resources": { "measured": true, "cpu_us_per_op": 1671.328, "cpu_seconds_total": 3.376083, "cpu_note": "User+system CPU across the timed loop, divided by iterations. On this harness the process is pinned to one core with nothing else scheduled, so this reads close to wall time by construction. It becomes informative only when utilisation varies under contention, which this harness deliberately does not create.", "max_rss_delta_bytes": 401408, "max_rss_bytes": 38088704, "rss_note": "Growth in the process high-water mark across the loop. ru_maxrss is a peak that never falls, so this is NOT a per-operation footprint and is not divided by the iteration count -- doing so would produce a confident, meaningless number.", "rss_unit_note": "ru_maxrss read as kilobytes (Linux convention) and converted to bytes." }, "note": "RSA key generation searches for primes, so its timing is both slow and unusually variable. It is sampled far fewer times than sign or verify; n_iterations records how many." }, "RSA-3072-PSS": { "scheme": "RSA-3072-PSS", "family": "RSA-PSS", "kind": "classical", "status": "ok", "keygen": { "mean_us": 174925.783, "median_us": 137919.628, "p95_us": 371545.869, "p99_us": 387320.488, "stdev_us": 118957.31, "min_us": 52543.904, "max_us": 391264.143, "ops_per_sec": 5.72, "n_iterations": 10, "ci95_low_us": 101196.632, "ci95_high_us": 248654.933, "std_error_us": 37617.6045, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 2569.925, "median_us": 2371.667, "p95_us": 3780.765, "p99_us": 4911.218, "stdev_us": 660.527, "min_us": 1965.483, "max_us": 10573.485, "ops_per_sec": 389.12, "n_iterations": 1000, "ci95_low_us": 2528.986, "ci95_high_us": 2610.864, "std_error_us": 20.8877, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 141.519, "median_us": 120.078, "p95_us": 302.106, "p99_us": 595.191, "stdev_us": 96.312, "min_us": 72.979, "max_us": 1326.116, "ops_per_sec": 7066.19, "n_iterations": 1000, "ci95_low_us": 135.55, "ci95_high_us": 147.488, "std_error_us": 3.0457, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 384, "public_key_bytes": 422, "parameters": { "hash": "SHA-256", "padding": "PSS", "salt_length": "digest length (32 B)" }, "classical_security_bits": 128, "security_source": "NIST SP 800-57 Part 1 Rev. 5, Table 2", "pairing_note": "No post-quantum counterpart is asserted. Pairing a classical scheme with a post-quantum one is a security-level claim rather than a measurement; the documented strength is published so the pairing can be argued rather than assumed.", "resources": { "measured": true, "cpu_us_per_op": 3623.351, "cpu_seconds_total": 7.282936, "cpu_note": "User+system CPU across the timed loop, divided by iterations. On this harness the process is pinned to one core with nothing else scheduled, so this reads close to wall time by construction. It becomes informative only when utilisation varies under contention, which this harness deliberately does not create.", "max_rss_delta_bytes": 0, "max_rss_bytes": 38240256, "rss_note": "Growth in the process high-water mark across the loop. ru_maxrss is a peak that never falls, so this is NOT a per-operation footprint and is not divided by the iteration count -- doing so would produce a confident, meaningless number.", "rss_unit_note": "ru_maxrss read as kilobytes (Linux convention) and converted to bytes." }, "note": "Same caveat as RSA-2048, more so: 3072-bit key generation is slow enough that a large sample would not fit the daily run's budget." }, "ECDSA-P256": { "scheme": "ECDSA-P256", "family": "ECDSA", "kind": "classical", "status": "ok", "keygen": { "mean_us": 20.167, "median_us": 17.779, "p95_us": 25.62, "p99_us": 87.9, "stdev_us": 15.394, "min_us": 17.537, "max_us": 281.098, "ops_per_sec": 49585.47, "n_iterations": 1000, "ci95_low_us": 19.213, "ci95_high_us": 21.121, "std_error_us": 0.4868, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 32.681, "median_us": 30.862, "p95_us": 40.24, "p99_us": 69.693, "stdev_us": 9.984, "min_us": 30.576, "max_us": 206.711, "ops_per_sec": 30598.57, "n_iterations": 1000, "ci95_low_us": 32.062, "ci95_high_us": 33.3, "std_error_us": 0.3157, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 113.214, "median_us": 91.862, "p95_us": 232.258, "p99_us": 427.114, "stdev_us": 82.313, "min_us": 88.832, "max_us": 1604.271, "ops_per_sec": 8832.84, "n_iterations": 1000, "ci95_low_us": 108.112, "ci95_high_us": 118.316, "std_error_us": 2.603, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 71, "public_key_bytes": 91, "parameters": { "hash": "SHA256", "encoding": "DER" }, "classical_security_bits": 128, "security_source": "NIST SP 800-57 Part 1 Rev. 5, Table 2", "pairing_note": "No post-quantum counterpart is asserted. Pairing a classical scheme with a post-quantum one is a security-level claim rather than a measurement; the documented strength is published so the pairing can be argued rather than assumed.", "resources": { "measured": true, "cpu_us_per_op": 82.343, "cpu_seconds_total": 0.247029, "cpu_note": "User+system CPU across the timed loop, divided by iterations. On this harness the process is pinned to one core with nothing else scheduled, so this reads close to wall time by construction. It becomes informative only when utilisation varies under contention, which this harness deliberately does not create.", "max_rss_delta_bytes": 409600, "max_rss_bytes": 39096320, "rss_note": "Growth in the process high-water mark across the loop. ru_maxrss is a peak that never falls, so this is NOT a per-operation footprint and is not divided by the iteration count -- doing so would produce a confident, meaningless number.", "rss_unit_note": "ru_maxrss read as kilobytes (Linux convention) and converted to bytes." }, "note": "Signature length varies between runs. ECDSA signatures are DER-encoded and the encoding is one or two bytes shorter when a component has a leading zero, so the size published here is measured, not the theoretical maximum." }, "ECDSA-P384": { "scheme": "ECDSA-P384", "family": "ECDSA", "kind": "classical", "status": "ok", "keygen": { "mean_us": 218.044, "median_us": 182.812, "p95_us": 420.564, "p99_us": 576.012, "stdev_us": 93.553, "min_us": 174.893, "max_us": 1138.81, "ops_per_sec": 4586.23, "n_iterations": 1000, "ci95_low_us": 212.246, "ci95_high_us": 223.843, "std_error_us": 2.9584, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "sign": { "mean_us": 295.391, "median_us": 255.241, "p95_us": 515.182, "p99_us": 752.055, "stdev_us": 115.502, "min_us": 244.33, "max_us": 1866.828, "ops_per_sec": 3385.34, "n_iterations": 1000, "ci95_low_us": 288.232, "ci95_high_us": 302.55, "std_error_us": 3.6525, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "verify": { "mean_us": 638.367, "median_us": 589.699, "p95_us": 863.401, "p99_us": 1165.153, "stdev_us": 121.405, "min_us": 572.13, "max_us": 1366.411, "ops_per_sec": 1566.5, "n_iterations": 1000, "ci95_low_us": 630.842, "ci95_high_us": 645.891, "std_error_us": 3.8392, "ci_note": "95% interval on the MEAN (normal approximation, z=1.96), not a prediction interval for a single operation. stdev_us describes how much individual samples scatter on this host; this describes how precisely the mean is known." }, "signature_bytes": 102, "public_key_bytes": 120, "parameters": { "hash": "SHA384", "encoding": "DER" }, "classical_security_bits": 192, "security_source": "NIST SP 800-57 Part 1 Rev. 5, Table 2", "pairing_note": "No post-quantum counterpart is asserted. Pairing a classical scheme with a post-quantum one is a security-level claim rather than a measurement; the documented strength is published so the pairing can be argued rather than assumed.", "resources": { "measured": true, "cpu_us_per_op": 511.826, "cpu_seconds_total": 1.535479, "cpu_note": "User+system CPU across the timed loop, divided by iterations. On this harness the process is pinned to one core with nothing else scheduled, so this reads close to wall time by construction. It becomes informative only when utilisation varies under contention, which this harness deliberately does not create.", "max_rss_delta_bytes": 0, "max_rss_bytes": 39096320, "rss_note": "Growth in the process high-water mark across the loop. ru_maxrss is a peak that never falls, so this is NOT a per-operation footprint and is not divided by the iteration count -- doing so would produce a confident, meaningless number.", "rss_unit_note": "ru_maxrss read as kilobytes (Linux convention) and converted to bytes." }, "note": "Same DER length variability as P-256." } } }